In a networked system, such as a networked computer system or a communications system, various entities are connected to a network. Communication between the entities can be established across the network. If problems occur with the network, such as failure of network infrastructure equipment, or overloading of the network, then this can cause degradation or loss of communication between entities communicating across the network. This can be costly for businesses who rely on these networks for the conduct of their day to day business. It is important that such problems are avoided, or rectified as quickly as possible.
An illustrative example of the above described problem arises in the operation of an IP telephony system. In an IP telephony system, TCP/IP enabled telephones are connected to a network. The network utilised can be an existing LAN which is already used for other tasks, such as networking a group of computers. This means that a telephone system can be installed in a premises that has a pre-existing network, and the telephones are configured to operate over the existing network. This allows a telephone system to be installed requiring significantly less wiring than a traditional telephone system, which requires wiring to be installed for the sole purpose of connecting the telephones to a local telephone exchange such as a PABX.
In an IP telephony system, calls made between telephones connected to the same LAN are established as communication paths between the telephones across the LAN. These communications paths are established by a system controller which controls operation of the system. This controller is typically in the form of a file server computer connected to the network. If a call is required to be made to a telephone not connected to the LAN, then the communication path may be routed across a WAN, or may be made via a gateway which establishes a connection to a traditional telephone network, such as a Public Switched Telephone Network (“PSTN”).
In a traditional telephone system, the originating telephone and the answering telephone remain connected by a permanent circuit for the duration of the call. The data transmission capability of the circuit is sufficient to carry a voice signal at an acceptable call quality. In contrast, during a telephone call in an IP telephony system, there exists no direct circuit connection between the originating and answering telephones. In the case of two telephones connected to the same network, the voice signal is encoded into data packets by one of the IP enabled telephones which are then transmitted across the network to the other telephone. The other telephone unencodes the received packets to reproduce the voice signal for the listener. This is a two way process to allow both call participants to speak and listen simultaneously.
In the case where the call is routed through a gateway to a PSTN, the data packets are transferred to or from the circuit switched PSTN by the gateway, which acts as an interface between the two systems.
The data packets, originating from either a gateway or an IP telephone, share the network with other data packets being sent and received by other entities connected to the network. These other entities can include other IP telephones, personal computers and other networked devices. During times of heavy network traffic, the packets sent and received by IP telephones may be held up due to bottlenecks caused by the overall level of network traffic. This can result in a depletion of the performance of the IP telephony system such as a drop in call quality experienced by the user, long delays in obtaining a dial tone, and even the dropping out of calls in progress. This disruption to the telephone service can be very costly in business. It is also frustrating to users who have come to expect trouble free telephone operation from traditional telephone systems.
Disruption to the system is similarly caused by delays in operation of the system controller. This could be due to various causes including a heavy demand on the controller, or that the controller has inadequate processing capacity to cope with normal demand.
Managing a networked system goes some way to avoiding, or assisting speedy resolution of, problems that occur with the system. These systems have been managed by reviewing activity logs created by entities connected to the network. However, these activity logs only give an indication of past operational activity and hence their usefulness in timely management of the network is limited.